A bias-corrected deep learning-based data assimilation method for riverine bathymetry inversion

The aim of this study was to evaluate the performance of using a deep learning-based method for measuring SBPs and DBPs and the effects of cuff inflation and deflation rates on the deep learning-based blood pressure (BP) measurement (in comparison with the manual auscultatory method). Forty healthy subjects were recruited. SBP and DBP were measured under four conditions (i.e. standard deflation, fast deflation, slow inflation and fast inflation) using both our newly developed deep learning-based method and the reference manual auscultatory method. The BPs measured under each condition were compared between the two methods. The performance of using the deep learning-based method to measure BP changes was also evaluated. There were no significant BP differences between the two methods (P > 0.05), except for the DBPs measured during the slow and fast inflation conditions. By applying the deep learning-based method, SBPs measured from fast deflation, slow inflation and fast inflation decreased significantly by 3.0, 3.5 and 4.7 mmHg (all P < 0.05), respectively, in comparison with the standard deflation condition. Whereas, corresponding DBPs measured from the slow and fast inflation conditions increased significantly by 5.0 and 6.8 mmHg, respectively (both P < 0.05). There were no significant differences in BP changes measured by the two methods in most cases (all P > 0.05, except for DBP change in the slow and fast inflation conditions). This study demonstrated that the deep learning-based method can achieve accurate BP measurement under the deflation and inflation conditions with different rates.

BackgroundSeveral deep learning-based methods have been proposed for addressing the long scanning time of magnetic resonance imaging. Most are trained using brain 3T magnetic resonance images, but is unclear whether performance is affected when applying these methods to different anatomical sites and at different field strengths.PurposeTo validate the denoising performance of deep learning-based reconstruction method trained by brain and knee 3T magnetic resonance images when applied to lumbar 1.5T magnetic resonance images.Material and MethodsUsing a 1.5T scanner, we obtained lumber T2-weighted sequences in 10 volunteers using three different scanning times: 228 s (standard), 119 s (double-fast), and 68 s (triple-fast). We compared the images obtained by the standard sequence with those obtained by the deep learning-based reconstruction-applied faster sequences.ResultsSignal-to-noise ratio values were significantly higher for deep learning-based reconstruction-double-fast than for standard and did not differ significantly between deep learning-based reconstruction-triple-fast and standard. Contrast-to-noise ratio values also did not differ significantly between deep learning-based reconstruction-triple-fast and standard. Qualitative scores for perceived signal-to-noise ratio and overall image quality were significantly higher for deep learning-based reconstruction-double fast and deep learning-based reconstruction-triple-fast than for standard. Average scores for sharpness, contrast, and structure visibility were equal to or higher for deep learning-based reconstruction-double-fast and deep learning-based reconstruction-triple-fast than for standard, but the differences were not statistically significant. The average scores for artifact were lower for deep learning-based reconstruction-double-fast and deep learning-based reconstruction-triple-fast than for standard, but the differences were not statistically significant.ConclusionThe deep learning-based reconstruction method trained by 3T brain and knee images may reduce the scanning time of 1.5T lumbar magnetic resonance images by one-third without sacrificing image quality.

Development and validation of a deep learning-based automatic auscultatory blood pressure measurement method

Develop and evaluate a deep learning-based automatic meningioma segmentation method for preoperative meningioma differentiation using radiomic features. A retrospective multicentre inclusion of MR examinations (T1/T2-weighted and contrast-enhanced T1-weighted imaging) was conducted. Data from centre 1 were allocated to training (n = 307, age = 50.94 ± 11.51) and internal testing (n = 238, age = 50.70 ± 12.72) cohorts, and data from centre 2 external testing cohort (n = 64, age = 48.45 ± 13.59). A modified attention U-Net was trained for meningioma segmentation. Segmentation accuracy was evaluated by five quantitative metrics. The agreement between radiomic features from manual and automatic segmentations was assessed using intra class correlation coefficient (ICC). After univariate and minimum-redundancy-maximum-relevance feature selection, L1-regularized logistic regression models for differentiating between low-grade (I) and high-grade (II and III) meningiomas were separately constructed using manual and automatic segmentations; their performances were evaluated using ROC analysis. Dice of meningioma segmentation for the internal testing cohort were 0.94 ± 0.04 and 0.91 ± 0.05 for tumour volumes in contrast-enhanced T1-weighted and T2-weighted images, respectively; those for the external testing cohort were 0.90 ± 0.07 and 0.88 ± 0.07. Features extracted using manual and automatic segmentations agreed well, for both the internal (ICC = 0.94, interquartile range: 0.88-0.97) and external (ICC = 0.90, interquartile range: 0.78-70.96) testing cohorts. AUC of radiomic model with automatic segmentation was comparable with that of the model with manual segmentation for both the internal (0.95 vs. 0.93, p = 0.176) and external (0.88 vs. 0.91, p = 0.419) testing cohorts. The developed deep learning-based segmentation method enables automatic and accurate extraction of meningioma from multiparametric MR images and can help deploy radiomics for preoperative meningioma differentiation in clinical practice. • A deep learning-based method was developed for automatic segmentation of meningioma from multiparametric MR images. • The automatic segmentation method enabled accurate extraction of meningiomas and yielded radiomic features that were highly consistent with those that were obtained using manual segmentation. • High-grade meningiomas were preoperatively differentiated from low-grade meningiomas using a radiomic model constructed on features from automatic segmentation.

The discovery of cancer subtypes based on unsupervised clustering helps in providing a precise diagnosis, guide treatment, and improve patients’ prognoses. Instead of single-omics data, multi-omics data can improve the clustering performance because it obtains a comprehensive landscape for understanding biological systems and mechanisms. However, heterogeneous data from multiple sources raises high complexity and different kinds of noise, which are detrimental to the extraction of clustering information. We propose an end-to-end deep learning based method, called Multi-omics Clustering Variational Autoencoders (MCluster-VAEs), that can extract cluster-friendly representations on multi-omics data. First, a unified network architecture with an attention mechanism was developed for accurately modeling multi-omics data. Then, using a novel objective function built from the Variational Bayes technique, the model was trained to effectively obtain the posterior estimation of the clustering assignments. Compared with 12 other state-of-the-art multi-omics clustering methods, MCluster-VAEs achieved an outstanding performance on benchmark datasets from the TCGA database. On the Pan Cancer dataset, MCluster-VAEs achieved an adjusted Rand index of approximately 0.78 for cancer category recognition, an increase of more than 18% compared with other methods. Furthermore, a survival analysis and clinical parameter enrichment tests conducted on 10 cancer datasets demonstrated that MCluster-VAEs provides comparable and even better results than many common integrative approaches. These results demonstrate that MCluster-VAEs are a powerful new tool for dissecting complex multi-omics relationships and providing new insights for cancer subtype discovery.

Recently, deep learning (DL) technology has been widely used in correspondence matching. The learning-based models are usually trained on benign image pairs with partial overlaps. Since DL model is usually data-dependent, non-overlapping images may be used as poison samples to fool the model and produce false registrations. In this study, we propose an outlier elimination-based assessment method (OEAM) to assess the registrations of learning-based correspondence matching method on partially overlapping and non-overlapping images. OEAM first eliminates outliers based on spatial paradox. Then OEAM implements registration assessment in two streams using the obtained core correspondence set. If the cardinality of the core set is sufficiently small, the input registration is assessed as a low-quality registration. Otherwise, it is assessed to be of high quality, and OEAM improves its registration performance using the core set. OEAM is a post-processing technique imposed on learning-based method. The comparison experiments are implemented on outdoor (YFCC100M) and indoor (SUN3D) datasets using four deep learning-based methods. The experimental results on registrations of partially overlapping images show that OEAM can reliably infer low-quality registrations and improve performance on high-quality registrations. The experiments on registrations of non-overlapping images demonstrate that learning-based methods are vulnerable to poisoning attacks launched by non-overlapping images, and OEAM is robust against poisoning attacks crafted by non-overlapping images.

Pairs Trading with Time-Series Deep Learning Models

The last few years have seen the development of numerous deep learning-based protein-ligand docking methods. They offer huge promise in terms of speed and accuracy. However, despite claims of state-of-the-art performance in terms of crystallographic root-mean-square deviation (RMSD), upon closer inspection, it has become apparent that they often produce physically implausible molecular structures. It is therefore not sufficient to evaluate these methods solely by RMSD to a native binding mode. It is vital, particularly for deep learning-based methods, that they are also evaluated on steric and energetic criteria. We present PoseBusters, a Python package that performs a series of standard quality checks using the well-established cheminformatics toolkit RDKit. The PoseBusters test suite validates chemical and geometric consistency of a ligand including its stereochemistry, and the physical plausibility of intra- and intermolecular measurements such as the planarity of aromatic rings, standard bond lengths, and protein-ligand clashes. Only methods that both pass these checks and predict native-like binding modes should be classed as having "state-of-the-art" performance. We use PoseBusters to compare five deep learning-based docking methods (DeepDock, DiffDock, EquiBind, TankBind, and Uni-Mol) and two well-established standard docking methods (AutoDock Vina and CCDC Gold) with and without an additional post-prediction energy minimisation step using a molecular mechanics force field. We show that both in terms of physical plausibility and the ability to generalise to examples that are distinct from the training data, no deep learning-based method yet outperforms classical docking tools. In addition, we find that molecular mechanics force fields contain docking-relevant physics missing from deep-learning methods. PoseBusters allows practitioners to assess docking and molecular generation methods and may inspire new inductive biases still required to improve deep learning-based methods, which will help drive the development of more accurate and more realistic predictions.

LiDAR is an active remote sensing technology widely used in forestry applications, such as forest resource surveys, tree information collection, and ecosystem monitoring. However, due to the resolution limitations of 3D-laser scanners and the canopy occlusion in forest environments, the tree point clouds obtained often have missing data. This can reduce the accuracy of individual tree segmentation, which subsequently affects the tree species classification. To address the issue, this study used point cloud data with RGB information collected by the UAV platform to improve tree species classification by completing the missing point clouds. Furthermore, the study also explored the effects of point cloud completion, feature selection, and classification methods on the results. Specifically, both a traditional geometric method and a deep learning-based method were used for point cloud completion, and their performance was compared. For the classification of tree species, five machine learning algorithms—Random Forest (RF), Support Vector Machine (SVM), Back Propagation Neural Network (BPNN), Quadratic Discriminant Analysis (QDA), and K-Nearest Neighbors (KNN)—were utilized. This study also ranked the importance of features to assess the impact of different algorithms and features on classification accuracy. The results showed that the deep learning-based completion method provided the best performance (avgCD = 6.14; avgF1 = 0.85), generating more complete point clouds than the traditional method. On the other hand, compared with SVM and BPNN, RF showed better performance in dealing with multi-classification tasks with limited training samples (OA-87.41%, Kappa-0.85). Among the six dominant tree species, Pinus koraiensis had the highest classification accuracy (93.75%), while that of Juglans mandshurica was the lowest (82.05%). In addition, the vegetation index and the tree structure parameter accounted for 50% and 30%, respectively, in the top 10 features in terms of feature importance. The point cloud intensity also had a high contribution to the classification results, indicating that the lidar point cloud data can also be used as an important basis for tree species classification.

Microscopic analyses of cytoskeleton organization are crucial for understanding various cellular activities, including cell proliferation and environmental responses in plants. Traditionally, assessments of cytoskeleton dynamics have been qualitative, relying on microscopy-assisted visual inspection. However, the transition to quantitative digital microscopy has introduced new technical challenges, with segmentation of cytoskeleton structures proving particularly demanding. In this study, we examined the utility of a deep learning-based segmentation method for accurate quantitative evaluation of cytoskeleton organization using confocal microscopic images of the cortical microtubules in tobacco BY-2 cells. The results showed that, although conventional methods sufficed for measurement of cytoskeleton angles and parallelness, the deep learning-based method significantly improved the accuracy of density measurements. To assess the versatility of the method, we extended our analysis to physiologically significant models in the context of changes in cytoskeleton density, namely Arabidopsis thaliana guard cells and zygotes. The deep learning-based method successfully improved the accuracy of cytoskeleton density measurements for quantitative evaluations of physiological changes in both stomatal movement in guard cells and intracellular polarization in elongating zygotes, confirming its utility in these applications. The results demonstrate the effectiveness of deep learning-based segmentation in providing precise and high-throughput measurements of cytoskeleton density, and has the potential to automate and expedite analyses of large-scale image datasets.

Optical coherence tomography (OCT) is a high-resolution non-invasive 3D imaging modality, which has been widely used for biomedical research and clinical studies. The presence of noise on OCT images is inevitable which will cause problems for post-image processing and diagnosis. The frame-averaging technique that acquires multiple OCT images at the same or adjacent locations can enhance the image quality significantly. Both conventional frame averaging methods and deep learning-based methods using averaged frames as ground truth have been reported. However, conventional averaging methods suffer from the limitation of long image acquisition time, while deep learning-based methods require complicated and tedious ground truth label preparation. In this work, we report a deep learning-based noise reduction method that does not require clean images as ground truth for model training. Three network structures, including Unet, super-resolution residual network (SRResNet), and our modified asymmetric convolution-SRResNet (AC-SRResNet), were trained and evaluated using signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR), edge preservation index (EPI) and computation time (CT). The effectiveness of these three trained models on OCT images of different samples and different systems was also investigated and confirmed. The SNR improvement for different sample images for L2-loss-trained Unet, SRResNet, and AC-SRResNet are 20.83 dB, 24.88 dB, and 22.19 dB, respectively. The SNR improvement for public images from different system for L1-loss-trained Unet, SRResNet, and AC-SRResNet are 19.36 dB, 20.11 dB, and 22.15 dB, respectively. AC-SRResNet and SRResNet demonstrate better denoising effect than Unet with longer computation time. AC-SRResNet demonstrates better edge preservation capability than SRResNet while Unet is close to AC-SRResNet. Eventually, we incorporated Unet, SRResNet, and AC-SRResNet into our graphic processing unit accelerated OCT imaging system for online noise reduction evaluation. Real-time noise reduction for OCT images with size of 512×512 pixels for Unet, SRResNet, and AC-SRResNet at 64 fps, 19 fps, and 17 fps were achieved respectively.

Network traffic classification has great significance for network security, network management and other fields. However, in recent years, the use of VPN and TLS encryption had presented network traffic classification with new challenges. Due to the great performances of deep learning in image recognition, many solutions have focused on the deep learning-based method and achieved positive results. A traffic classification method based on deep learning is provided in this paper, where the concept of Packet Block is proposed, which is the aggregation of continuous packets in the same direction. The features of Packet Block are extracted from network traffic, and then transformed into images. Finally, convolutional neural networks are used to identify the application type of network traffic. The experiment is conducted using captured OpenVPN dataset and public ISCX-Tor dataset. The results shows that the accuracy is 97.20% in OpenVPN dataset and 93.31% in ISCX-Tor dataset, which is higher than the state-of-the-art methods. This suggests that our approach has the ability to meet the challenges of VPN and TLS encryption.

In the medical imaging domain, deep learning-based methods have yet to see widespread clinical adoption, in part due to limited generalization performance across different imaging devices and acquisition protocols. The deviation between estimated brain age and biological age is an established biomarker of brain health and such models may benefit from increased cross-site generalizability. To develop and evaluate a deep learning-based image harmonization method to improve cross-site generalizability of deep learning age prediction. Retrospective. Eight thousand eight hundred and seventy-six subjects from six sites. Harmonization models were trained using all subjects. Age prediction models were trained using 2739 subjects from a single site and tested using the remaining 6137 subjects from various other sites. Brain imaging with magnetization prepared rapid acquisition with gradient echo or spoiled gradient echo sequences at 1.5 T and 3 T. StarGAN v2, was used to perform a canonical mapping from diverse datasets to a reference domain to reduce site-based variation while preserving semantic information. Generalization performance of deep learning age prediction was evaluated using harmonized, histogram matched, and unharmonized data. Mean absolute error (MAE) and Pearson correlation between estimated age and biological age quantified the performance of the age prediction model. Our results indicated a substantial improvement in age prediction in out-of-sample data, with the overall MAE improving from 15.81 (±0.21) years to 11.86 (±0.11) with histogram matching to 7.21 (±0.22) years with generative adversarial network (GAN)-based harmonization. In the multisite case, across the 5 out-of-sample sites, MAE improved from 9.78 (±6.69) years to 7.74 (±3.03) years with histogram normalization to 5.32 (±4.07) years with GAN-based harmonization. While further research is needed, GAN-based medical image harmonization appears to be a promising tool for improving cross-site deep learning generalization. 4 TECHNICAL EFFICACY: Stage 1.

Accurate deformable four-dimensional (4D) (three-dimensional in space and time) medical images registration is essential in a variety of medical applications. Deep learning-based methods have recently gained popularity in this area for the significantly lower inference time. However, they suffer from drawbacks of non-optimal accuracy and the requirement of a large amount of training data. A new method named GroupRegNet is proposed to address both limitations. The deformation fields to warp all images in the group into a common template is obtained through one-shot learning. The use of the implicit template reduces bias and accumulated error associated with the specified reference image. The one-shot learning strategy is similar to the conventional iterative optimization method but the motion model and parameters are replaced with a convolutional neural network and the weights of the network. GroupRegNet also features a simpler network design and a more straightforward registration process, which eliminates the need to break up the input image into patches. The proposed method was quantitatively evaluated on two public respiratory-binned 4D-computed tomography datasets. The results suggest that GroupRegNet outperforms the latest published deep learning-based methods and is comparable to the top conventional method pTVreg. To facilitate future research, the source code is available at https://github.com/vincentme/GroupRegNet.

With the increasing frequency and severity of disasters and accidents, there is a growing need for efficient emergency alert systems. The ultra-high definition (UHD) broadcasting service based on Advanced Television Systems Committee (ATSC) 3.0, a leading terrestrial digital broadcasting system, offers such capabilities, including a wake-up function for minimizing damage through early alerts. In case of a disaster situation, the emergency alert wake-up signal is transmitted, allowing UHD TVs to be activated, enabling individuals to receive emergency alerts and access emergency broadcasting content. However, conventional methods for detecting the bootstrap signal, essential for this function, typically require an ATSC 3.0 demodulator. In this paper, we propose a novel deep learning-based method capable of detecting an emergency wake-up signal without the need for an ATSC 3.0. The proposed method leverages deep learning techniques, specifically a deep neural network (DNN) structure for bootstrap detection and a convolutional neural network (CNN) structure for wake-up signal demodulation and to detect the bootstrap and 2 bit emergency alert wake-up signal. Specifically, our method eliminates the need for Fast Fourier Transform (FFT), frequency synchronization, and interleaving processes typically required by a demodulator. By applying a deep learning in the time domain, we simplify the detection process, allowing for the detection of an emergency alert signal without the full suite of demodulator components required for ATSC 3.0. Furthermore, we have verified the performance of the deep learning-based method using ATSC 3.0-based RF signals and a commercial Software-Defined Radio (SDR) platform in a real environment.